codeburn/README.md

CodeBurn

See where your AI coding tokens go.

By task type, tool, model, MCP server, and project. Tracks one-shot success rate per activity type so you can see where the AI nails it first try vs. burns tokens on edit/test/fix retries. Interactive TUI dashboard with gradient charts, responsive panels, and keyboard navigation. macOS menu bar widget via SwiftBar. CSV/JSON export.

Works by reading Claude Code session transcripts directly from disk. No wrapper, no proxy, no API keys. Pricing from LiteLLM (auto-cached, all models supported).

Install

npm install -g codeburn

Or run without installing:

npx codeburn

Requirements

• Node.js 18+
• Claude Code (reads ~/.claude/projects/ session data)

Usage

codeburn                    # interactive dashboard (default: 7 days)
codeburn today              # today's usage
codeburn month              # this month's usage
codeburn report -p month    # same as above
codeburn status             # compact one-liner (today + month)
codeburn status --format json
codeburn export             # CSV with today, 7 days, 30 days
codeburn export -f json     # JSON export

Arrow keys switch between Today / 7 Days / Month. Press q to quit, 1 2 3 as shortcuts.

Menu Bar

codeburn install-menubar    # install SwiftBar/xbar plugin
codeburn uninstall-menubar  # remove it

Requires SwiftBar (brew install --cask swiftbar). Shows today's cost in the menu bar with a flame icon. Dropdown shows activity breakdown, model costs, and token stats for today, 7 days, and month. Refreshes every 5 minutes.

What it tracks

13 task categories classified from tool usage patterns and user message keywords. No LLM calls, fully deterministic.

Category	What triggers it
Coding	Edit, Write tools
Debugging	Error/fix keywords + tool usage
Feature Dev	"add", "create", "implement" keywords
Refactoring	"refactor", "rename", "simplify"
Testing	pytest, vitest, jest in Bash
Exploration	Read, Grep, WebSearch without edits
Planning	EnterPlanMode, TaskCreate tools
Delegation	Agent tool spawns
Git Ops	git push/commit/merge in Bash
Build/Deploy	npm build, docker, pm2
Brainstorming	"brainstorm", "what if", "design"
Conversation	No tools, pure text exchange
General	Skill tool, uncategorized

Breakdowns: daily cost chart, per-project, per-model (Opus/Sonnet/Haiku/GPT-4o/Gemini), per-activity with one-shot rate, core tools, MCP servers.

One-shot rate: For categories that involve code edits, CodeBurn detects edit/test/fix retry cycles (Edit -> Bash -> Edit patterns). The 1-shot column shows the percentage of edit turns that succeeded without retries. Coding at 90% means the AI got it right first try 9 out of 10 times.

Pricing: Fetched from LiteLLM model prices (auto-cached 24h at ~/.cache/codeburn/). Handles input, output, cache write, cache read, and web search costs. Fast mode multiplier for Claude. Fallback to hardcoded prices if fetch fails.

How it reads data

Claude Code stores session transcripts as JSONL at ~/.claude/projects/<sanitized-path>/<session-id>.jsonl. Each assistant entry contains model name, token usage (input, output, cache read, cache write), tool_use blocks, and timestamps.

CodeBurn reads these files, deduplicates messages by API message ID (prevents double-counting across sessions), filters by date range per entry (not per session), and classifies each turn.

Project structure

src/
  cli.ts          Commander.js entry point
  dashboard.tsx   Ink TUI (React for terminals)
  parser.ts       JSONL reader, dedup, date filter
  models.ts       LiteLLM pricing, cost calculation
  classifier.ts   13-category task classifier
  types.ts        Type definitions
  format.ts       Text rendering (status bar)
  menubar.ts      SwiftBar plugin generator
  export.ts       CSV/JSON multi-period export

License

MIT

Credits

Inspired by ccusage. Pricing data from LiteLLM.

Built by AgentSeal.